The second moment of cubic Dirichlet L-functions over function fields
Shivani Goel, Anwesh Ray · 2025
Read The second moment of cubic Dirichlet L-functions over function fields by Shivani Goel, Anwesh Ray (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-12015.txt.
Introduction
Shivani Goel and Anwesh Ray average a family of L-functions in a 2025 mathematics paper. L-functions encode deep arithmetic information, and their values at the central point carry conjectural weight. The authors work over function fields, where the rational functions over a finite field replace the integers, and study cubic Dirichlet L-functions there.
They compute the second moment, the average of the squared central values, for primes whose size obeys a congruence condition. Earlier work by other authors had settled the first moment in this setting. The proof adapts classic analytic tools, Perron's formula and the approximate functional equation, to the function-field world. A key step converts the problem into averages of Gauss sums, estimated using known results. Moment estimates like these test random-matrix predictions about L-function families.
Catalog record
| CREATOR | Shivani Goel, Anwesh Ray |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 181 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-12015
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
The second moment of cubic Dirichlet L-functions over function fields Shivani Goel, Anwesh Ray Abstract . In this article, we study the second moment of cubic Dirichlet L-functions at the central point s = 1 / 2 s=1/2 over the rational function field 𝔽 q ( T ) \mathbb{F}_{q}(T) , where q q is a power of an odd prime satisfying q ≡ 2 ( mod 3 ) q\equiv 2\pmod{3} . Our result extends prior work of David, Florea and Lalin [ DFL22 ] , who obtained an asymptotic formula for the first moment. Our approach relies on analytic techniques (Perron’s formula, approximate functional equation, etc), adapted to the function field context. A key step in the construction is to relate second moment to certain averages of Gauss sums, which are estimated in loc. cit. using results of Kubota [ Kub69 ] and Hoffstein [ Hof92 ] . 2020 Mathematics Subject Classification 11R59, 11M38, 11R58 1. Introduction 1.1. Background and motivation The study of L-functions in the function field setting has long served as a fertile testing ground for deep conjectures in number theory and arithmetic geometry. The arithmetic of the polynomial ring \mathbb{F}_{q}[T] , where q is a power of a fixed odd prime p , closely parallels the classical theory of the integers, with zeta and L -functions playing an analogous role. In this setting, one gains access to powerful tools from algebraic geometry, most notably through the interpretation of L -functions as generating series encoding the action of the Frobenius endomorphism on the étale cohomology of algebraic curves over finite fields. For a systematic introduction to the topic, the reader may consult [ Ros02 , KW01 , Tha04 ] . The study of moments of Dirichlet L -functions at the central point \tfrac{1}{2} is a fundamental problem in analytic number theory. For quadratic characters \chi over the rationals, the first and second moments of L(\tfrac{1}{2},\chi) were computed by Jutila [ Jut81 ] , who established the asymptotic formula \sum_{0<d\leqslant D}L\left(\tfrac{1}{2},\chi_{d}\right)=\frac{P(1)}{4\zeta(2)}D\left\{\log\left(\frac{D}{\pi}\right)+\frac{\Gamma^{\prime}}{\Gamma}\left(\tfrac{1}{4}\right)+4\gamma-1+4\frac{P^{\prime}}{P}(1)\right\}+O\left(D^{3/4+\varepsilon}\right), where d ranges over squarefree numbers between 1 and D , and P(s)\mathrel{\mathop{\mathchar 58\relax}}=\prod_{p}\left(1-\frac{1}{(p+1)p^{s}}\right) . Moreover, for the second moment, he showed that \sum_{0<d\leqslant D}L\left(\tfrac{1}{2},\chi_{d}\right)^{2}=\frac{c_{1}}{\zeta(2)}D\log^{3}D+O\left(D(\log D)^{5/2+\varepsilon}\right), for an explicit constant c_{1}>0 . Related results for the third moment were obtained by Soundararajan [ Sou00 ] . There are conjectural asymptotics for moments of quadratic Dirichlet L -functions over \mathbb{Q} due to Keating and Snaith [ KS00 ] based on random matrix theory. Baier and Young [ BY10 ] obtained the smoothed first moment for cubic Dirichlet characters over \mathbb{Q} . Certain results have also been obtained over \mathbb{Q} for thin families of characters by Luo [ Luo04 ] , Friedberg–Hoffstein–Lieman [ FHL03 ] . More recently, David, de Faveri, Dunn, and Stucky [ DdFDS24 ] have computed the first and second moments of L(1/2,\chi_{q}) for primitive cubic characters \chi_{q} with q\in\mathbb{Z}[\mu_{3}] satisfying q\equiv 1\pmod{9} , thereby establishing an unconditional positive proportion of non-vanishing at the central point. Over the rational function field, the first results of this nature are due to Hoffstein and Rosen [ HR92 ] , who obtained an asymptotic formula for the first moment of L(\frac{1}{2},\chi) where \chi ranges over quadratic Dirichlet characters of the rational function field \mathbb{F}_{q}(T) . Such L -functions are closely related to the zeta functions of hyperelliptic curves over \mathbb{F}_{q} . Their methods involve the use of Eisenstein series for the metaplectic 2 -fold cover of \operatorname{GL}_{2} . A different proof of this result, using classical methods was obtained by Andrade and Keating [ AK12 ] . This classical approach was further extended by Florea, who obtained asymptotic formulae for the second, third and fourth moments [ Flo17b , Flo17a ] . 1.2. Main results For cubic characters over the rational function field \mathbb{F}_{q}(T) , an asymptotic formula for the first moment was obtained by David–Florea–Lalin [ DFL22 ] . Let L_{q}(s,\chi) denote the L -function associated to a character \chi of \mathbb{F}_{q}(T) . A natural dichotomy arises depending on whether q\equiv 1\pmod{3} or q\equiv 2\pmod{3} , since in the former case, the third roots of unity are contained in \mathbb{F}_{q}^{\times} . Let \mu_{3} (resp. \mu_{3,q} ) denote the group of third roots of unity in \mathbb{C}^{\times} (resp. in \mathbb{F}_{q}^{\times} ), and fix a bijection \Omega\mathrel{\mathop{\mathchar 58\relax}}\mu_{3}\xrightarrow{\sim}\mu_{3,q} . Define a character \chi_{3} on \mathbb{F}_{q}^{\times} by setting \chi_{3}(a)\mathrel{\mathop{\mathchar 58\relax}}=\Omega^{-1}(a^{\frac{q-1}{3}}) . One refers to the case q\equiv 1\pmod{3} as the Kummer case , and to q\equiv 2\pmod{3} as the non-Kummer case . When q\equiv 2\pmod{3} , Theorem 1.1 of loc. cit. asserts that: \sum_{\begin{subarray}{c}\chi\text{ primitive cubic}\\ \mathrm{genus}(\chi)=g\end{subarray}}L_{q}(1/2,\chi)=\frac{\zeta_{q}(3/2)}{\zeta_{q}(3)}C(q)q^{g+2}+O(q^{\frac{7g}{8}+\varepsilon g}), where C(q) is an explicit constant depending only on q and \varepsilon>0 . On the other hand when q\equiv 1\pmod{3} , their result applies to primitive cubic characters \chi satisfying the condition \chi|_{\mathbb{F}_{q}^{\times}}=\chi_{3} . Theorem 1.2 of loc. cit. states that there exist explicit constants C_{1}(q),C_{2}(q)>0 such that for \varepsilon>0 , \sum_{\begin{subarray}{c}\chi\text{ primitive cubic}\\ \mathrm{genus}(\chi)=g\\ \chi|_{\mathbb{F}_{q}^{\times}}=\chi_{3}\end{subarray}}L_{q}(1/2,\chi)=C_{1}(q)gq^{g}+C_{2}(q)q^{g}+O\left(q^{g\frac{1+\sqrt{7}}{4}+\varepsilon g}\right). The hypothesis that \chi restricts to \chi_{3} on \mathbb{F}_{q}^{\times} serves to simplify the analysis by ensuring that all L -functions in the family satisfy a uniform functional equation which is easier to work with. Restrictions of this nature are common in the field, and help simplify many calculations. In [ AK12 , Flo17b , Flo17a ] , the assumption that q\equiv 1\pmod{4} is made. Furthermore, in the latter two works, q is assumed to be a prime. In this article we consider the non-Kummer case, i.e., we assume that q is an odd prime power and q\equiv 2\pmod{3} . We obtain an asymptotic for the second moment of L_{q}(\frac{1}{2},\chi) . Theorem A . With respect to notation above, assume that q\equiv 2\pmod{3} , then, (1.1) \sum_{\begin{subarray}{c}\chi\text{ primitive cubic}\\ \mathrm{genus}(\chi)=g\end{subarray}}L_{q}(1/2,\chi)^{2}=\frac{g(g+2)A_{q}(\frac{1}{q^{2}},\frac{1}{q^{3/2}})\zeta_{q}(3/2)^{2}}{8\zeta_{q}(3)}q^{g+2}+O(q^{g}), where A_{q} is the function defined in ( 3.9 ). The case q\equiv 1\pmod{3} is not addressed in this paper, as it necessitates the analysis of certain twisted averages of Gauss sums, which have not been studied in the literature. The difficulties arising in this case are substantial and will be addressed in a more general framework in forthcoming work. 1.3. Methodology Let us briefly outline the strategy of the proof of Theorem A . For any positive integer n , let \mathcal{M}_{q,n}\subset\mathbb{F}_{q}[T] be the set of all monic polynomials of degree n . Assume henceforth that q\equiv 2\pmod{3} . Let \mathcal{M}_{q}\subset\mathbb{F}_{q}[T] denote the set of all monic polynomials, and let \mathcal{M}_{q,\leq n}\subset\mathcal{M}_{q} denote the subset of monic polynomials of degree \leq n . Define \mathcal{H}_{q,n} as the subset of \mathcal{M}_{q,n} consisting of squarefree monic polynomials. It is easy to see that \#\mathcal{M}_{q,n}=q^{n},\quad\#\mathcal{M}_{q,\leq n}=1+q+\dots+q^{n}=\frac{q^{n+1}-1}{q-1},\quad\#\mathcal{H}_{q,n}=q^{n}\left(1-\frac{1}{q}\right). Let \chi be a cubic character and L_{q}(s,\chi) be the associated L -function over \mathbb{F}_{q}[T] . Much like number field case, L_{q}(s,\chi) satisfies a functional equation (see Lemma 2.2 ), which yields the approximate functional equation : (1.2) \begin{split}L_{q}(1/2,\chi)^{2}=\left(1-\frac{1}{\sqrt{q}}\right)^{2}\Bigg(&\sum_{i=0}^{A}(i+1)\sum_{f\in\mathcal{M}_{\leq A-i}}\frac{\chi(f)\,d(f)}{\sqrt{|f|}}\\ &+\sum_{i=0}^{2g-A-1}(i+1)\,\omega(\chi)^{2}\sum_{f\in\mathcal{M}_{\leq 2g-A-1-i}}\frac{\bar{\chi}(f)\,d(f)}{\sqrt{|f|}}\Bigg).\end{split} Here, d(f) is the divisor function, \omega(\chi) is the sign of \chi and the choice of A is postponed till the end of the argument. We refer to Proposition 2.3 for further details. Associated to a polynomial F\in\mathcal{H}_{q^{2},g/2+1} , we have a primitive cubic Dirichlet character \chi_{F} defined in section 2.5 . The second moment of central L-values over primitive cubic characters of genus g is then given by \sum_{\begin{subarray}{c}\chi\text{